Synthesis and characterization of constrained peptidomimetic dipeptidyl peptidase IV inhibitors: Amino-lactam boroalanines

Article abstract of DOI:10.1021/jm061321+

We describe here the epimerization-free synthesis and characterization of a new class of conformationally constrained lactam aminoboronic acid inhibitors of dipeptidyl peptidase IV (DPP IV; E.C. 3.4.14.5). These compounds have the advantage that they cann

Full text of DOI:10.1021/jm061321+

J. Med. Chem. 2007, 50, 2391-2398
2391
Synthesis and Characterization of Constrained Peptidomimetic Dipeptidyl Peptidase IV
Inhibitors: Amino-Lactam boroAlanines
Jack H. Lai, Wengen Wu, Yuhong Zhou, Hlaing H. Maw, Yuxin Liu, Lawrence J. Milo, Jr., Sarah E. Poplawski,†
Gillian D. Henry, James L. Sudmeier, David G. Sanford, and William W. Bachovchin*
Contribution from the Department of Biochemistry, Tufts UniVersity School of Medicine, 136 Harrison AVenue, Boston, Massachusetts 02111
ReceiVed NoVember 14, 2006
We describe here the epimerization-free synthesis and characterization of a new class of conformationally
constrained lactam aminoboronic acid inhibitors of dipeptidyl peptidase IV (DPP IV; E.C. 3.4.14.5). These
compounds have the advantage that they cannot undergo the pH-dependent cyclization prevalent in most
dipeptidyl boronic acids that attenuates their potency at physiological pH. For example, D-3-amino-1-[L-1-
boronic-ethyl]-pyrrolidine-2-one (amino-D-lactam-L-boroAla), one of the best lactam inhibitors of DPP IV,
is several orders of magnitude less potent than L-Ala-L-boroPro, as measured by Ki values (2.3 nM vs 30
pM, respectively). At physiological pH, however, it is actually more potent than L-Ala-L-boroPro, as measured
by IC50 values (4.2 nM vs 1400 nM), owing to the absence of the potency-attenuating cyclization. In an
interesting and at first sight surprising reversal of the relationship between stereochemistry and potency
observed with the conformationally unrestrained Xaa-boroPro class of inhibitors, the L-L diastereomers of
the lactams are orders of magnitude less effective than the D-L lactams. However, this interesting reversal
and the unexpected potency of the D-L lactams as DPP IV inhibitors can be understood in structural terms,
which is explained and discussed here.
Introduction
Dipeptidyl peptidase IV (DPP IV; E.C. 3.4.14.5) is a serine
protease found on the surface of nearly every organ and tissue
in the body.1 It is strongly expressed on the surface of T cells,
Figure 1. pH-Dependent equilibrium between linear and cyclic forms
where it is also known as CD26, a T cell activation antigen.2
of Xaa-boroPro. The open chain form (left) is favored at low pH values,
But it also occurs in soluble form in plasma and other body
and the cyclic form (right) is favored at high pH.
fluids.3 As an enzyme, DPP IV cleaves dipeptides from the
amino terminus of peptides and proteins containing either
physiological pH. The attenuation can be quite large. For L-Ala-
L-proline or L-alanine at the penultimate position.4 Such removal
L-boroPro, for example, it is several 1000-fold.13
of a dipeptide from the N-terminus of the incretin hormones
Cyclization can be eliminated by N-terminal substitutions that
glucagon-like peptide-1 (GLP-1) and glucose-dependent insuli-
render the unshared electron pair on the N-terminal nitrogen
notropic peptide (GIP) eliminates their biological activities.5
unavailable for nucleophilic attack on the boron, such as
Because GLP-1 and GIP have multiple activities, all of which
acylation or trimethylation. However, such substituents also
strongly reduce inhibitor potency toward DPP IV8,14 because
they render the electron pair unavailable for protonation, and a
protonated amino group is required for binding of both substrates
and inhibitors. N-Alkylations that leave the unshared pair
available for protonation can reduce the tendency to cyclize,
largely through steric hindrance, but also often reduce inhibitory
potency as well.15
In an effort to prevent the intramolecular N-B cyclization,
while retaining an unsubstituted N-terminal amino group, we
synthesized two stereospecifically constrained peptidomimetics
amino-D-lactam-L-boroAla (1) and amino-L-lactam-L-boroAla
(2), as boronic acid analogs to Freidinger’s five-membered
lactam-bridged dipeptide scaffold.16
As depicted in Figure 2, these lactam analogs are derived
from the previously reported DPP IV inhibitor L-Ala-L-boroPro
(3)13 and its epimer D-Ala-L-boroPro (4). Most notably, the
cyclic pyrrolidine moiety at P1 is replaced by an acyclic amino
acid (alanine) analog. In these lactam analogs, by breaking apart
atoms C5 and C6 and connecting C4 and C5, ω is locked into
the trans conformation (∼ 180°), the conformation that binds
to the enzyme. Furthermore, the angle ψ becomes part of the
five-membered lactam ring, and inspection of the molecular
models reveals that it is restricted to 120 ( 30o in compound 1
would be highly desirable attributes in an antidiabetogenic drug,
and because DPP IV has been shown to be largely responsible
for the very short half-live of these hormones in vivo,6 inhibitors
of DPP IV are of substantial interest for their therapeutic
potential in the treatment of type II diabetes.7
Dipeptide proline boronic acids of the type Xaa-boroPro
(-boroPro refers to proline in which the C-terminal carboxylate
has been replaced by a boronyl group and Xaa is any unblocked
amino acid) are among the most potent inhibitors of DPP IV
known.8-10 For example, L-Pro-L-boroPro has a Kia value of
16 pM for DPP IV.9 However, these inhibitors undergo a
reversible, pH-dependent intramolecular cyclization between the
N-terminal amine and the C-terminal boron (Figure 1).11,12 The
cyclic structure, favored at high pH, is devoid of inhibitory
activity. Thus, the net effect of the pH-dependent cyclization
reaction is to attenuate the potency of these inhibitors at
* To whom correspondence should be addressed. Phone: 617-636-6881.
Fax: 617-636-2409. E-mail: william.bachovchin@tufts.edu.
† Present address: Department of Neurology, Brigham and Women’s
Hospital, 75 Francis Street, Boston, Massachusetts 02115.
a Abbreviations: HSQC, heteronuclear single quantum coherence;
NOESY, nuclear Overhauser spectroscopy; Ki, inhibition constant; IC50,
median inhibition concentration.
10.1021/jm061321+ CCC: $37.00 © 2007 American Chemical Society
Published on Web 04/26/2007

2392 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 10
Lai et al.
via palladium-catalyzed hydrogenation. The hydroxyl group of
10 was mesylated by MsCl and cyclized with the amide
nitrogen, followed by treatment with 1 equiv of LDA to form
the pyrrolidinone boronic ester.21 Following the workup, the
crude product was deprotected using BCl3, which gave a mixture
of the HCl salts of pyrrolidinoneboronic acids 1 and 2. The
combined yield of the final three synthetic steps was 46%, with
the ratio of 1:2 ranging from 1:4 to 1:19.
Compound 1 was first prepared from the commercially
available N-Boc-D-methionine (11), following the strategy
depicted in Scheme 2. However, the ratio of diastereomers 1 to
2 was 1.5:1, as determined by 1H NMR. In addition, the
cyclopropyl-bearing byproduct (12), resulting from the intramo-
lecular C-alkylations of the corresponding methionine sulfonium
salts, was also isolated. The ratio of these isolates (1:2:12) was
9:6:5.
Figure 2. Design of the γ-lactam boronate derivatives. The cyclic
pyrrolidine moiety in 3 and 4 is replaced by the corresponding acyclic
amino acid 1 and 2.
and -120 ( 30° in 2 (the flexibility of (30° or somewhat more
is due to the puckering of five-membered rings).
To enhance the stereopurity of 1, we used the unique
properties of the 9-phenylfluorenyl (Pf) protecting group.22 The
steric bulk of the Pf group attached to the amine of an optically
pure methionine prevents deprotonation at CR by strong bases
and results in proton abstraction exclusively at the amide
nitrogen. Introduction of the Pf group involved conversion from
the pre-existing Boc group. This was accomplished by treating
13 with 4 N HCl in dioxane, followed by protection, using N-Pf
formation conditions23 to afford N(Pf)-Met-boroAla-pn (14) in
a yield of 79% (Scheme 3). Compound 14 was subsequently
converted to the sulfonium salt by reaction with methyl iodide.
Formation of the lactam ring was induced by sodium hydride
in 1:1 DMF/methylene chloride. Deprotection of the crude
dipeptide derivative by BCl3, followed by HPLC purification,
gave the desired pyrrolidinoneboronic acid (1) as a hydrogen
chloride salt (24% overall yield in three steps). The final product
We report here the success of this backbone-constraining
strategy in the epimerization-free synthesis and characterization
of a noncyclizing boronic acid DPP IV inhibitor of low
nanomolar potency. The paradoxical identity of the active epimer
has led to increased insight for the design of DPP IV inhibitors.
Results and Discussion
Chemistry. Table 1 summarizes the compounds synthesized
in this study. Compounds 3-6 were resynthesized, following
the procedures previously reported by the authors.13
The key to the synthesis of the target amino-γ-lactam
boronates (1,2) involves use of an appropriate leaving group to
aid intramolecular nitrogen-carbon bond formation upon treat-
ment with a strong base. We investigated various synthetic
routes starting from protected chiral dipeptides for the synthesis
of each γ-lactam derivative, however, several resulted in
epimerization at the P2 R-carbon.
1
was verified by H NMR, which indicated that epimerization
was successfully avoided. The same method was used to give
the pure L-diastereoisomer (2; 16% total yield in five steps),
starting with L-N-Boc-Met-L-BoroAla-pn (8; Scheme 3).
A Boc-protected dipeptide derivative containing a side chain
sulfonium salt16 was used in the synthesis of 1 and 2. This
enabled lactam ring formation through intramolecular alkylation
of the amide nitrogen. As shown in Scheme 1 (route A),
commercially available N-Boc-L-methionine (7) was first con-
densed with L-boroAla-pn hydrochloride18 under standard
coupling conditions19 to afford the protected methionine dipep-
tide (8) in 88% yield. Treatment of 8 with methyl iodide
followed by sodium hydride in 1:1 DMF/DCM gave crude
protected dipeptide lactam. Without further purification of the
crude mixture, concurrent removal of the Boc and pinane groups
by BCl3 gave a 1:19 diastereomeric ratio of 1 and 2, as revealed
Finally, formation of an undesired structural isomer, azetidine
(15), via terminal N-alkylation (path B, Scheme 4) instead of
the desired lactam (2), via amide N-alkylation (path A, Scheme
4), was a concern throughout the synthesis of the target lactam
derivative (2; the same concern was applied to 1). To confirm
this, we synthesized 15 via a three-step synthetic procedure
starting with L-azetidine-2-carboxylic and showed that it has
different characteristics than 2 in HPLC retention time and 1H
1D and 2D NOESY NMR spectra. Confirming evidence arises
from the biological assay with DPP IV: compound 2 is about
7-fold less potent than 15 and exhibits a pH-independent
inhibition profile, in contrast to the pH-dependent profile of 15.
1
by H NMR analysis, in 41% yield. Presumably, the minor
product (1) resulted from the epimerization of the P2 R-proton
by use of the strong base used in the lactam ring formation
step. Replacement of sodium hydride with either a stoichiometric
or a smaller amount of the N-methylacetamide lithium salt20
did not reduce the amount of 1 formed. Although 1 and 2 are
both desired targets of this study, it was difficult to separate
them via HPLC due to their similar retention times and high
polarities. Attempts to purify the corresponding protected
precursors by HPLC or silica gel chromatography also failed.
To reduce the amount of epimerization resulting from the
intramolecular alkylation reaction in the synthesis of 2, a
different leaving group was sought (route B of Scheme 1). The
hydroxyl group of the P2 side chain was converted to the
mesylated alcohol to aid in the intramolecular nitrogen-carbon
bond formation. Following a standard peptide synthesis protocol,
commercially available N-Boc-O-benzyl-L-homoserine (9) was
coupled to L-boroAla-pn using HATU. Compound 10 was
obtained in 53% yield upon deprotection of the O-benzyl group
Enzyme Assay. Our previous studies of Xaa-boroPro dipep-
tides showed that DPP IV inhibitors exist predominantly in the
open, trans form of proline at low pH values (e.g., 2) and in
the cyclic, cis form at high pH (e.g., 8).11,12 Half-lives for
cyclization and uncyclization of these boroPro inhibitors are
typically measured in hours,12 largely due to the high-energy
barrier of proline cis-trans interconversion. Because assays can
be performed in a matter of minutes, and the cyclic structures
are noninhibitory, the ratio of the cyclic (inactive) to open chain
(active) material can be easily approximated using simple
enzyme inhibition assays. The inhibitor is incubated in solution
of a given pH (typically pH 8.0 and 2.0), long enough (generally
24 h) to allow the open and cyclic structures to reach
equilibrium. The inhibitors are then added to an assay solution
at pH 8.0 containing enzyme, and incubated for 10 min before
substrate is added and the enzyme activity is measured. Ten

Constrained Lactam Boronate DPP IV Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 10 2393
Table 1. Summary of Boronic Acids Synthesized, Torsional Angles Calculated by Ab Initio Methods, and Their Inhibition Parameters vs DPP IV
b
b
ψ(P2)
(deg)
ω
(deg)
φ(P1)
(deg)
Ki
(nM)
IC50
IC50
cmpd
namea
(pH 2, nM)
(pH 8, nM)
CIc
0.8
1.0
2600
1d
2d
3e
4e
5e
6e
amino-D-lactam-L-boroAla
amino-L-lactam-L-boroAla
L-Ala-L-boroPro
D-Ala-L-boroPro
L-Val-L-boroPro
+146.3
-106.5
+155.4
-158.0
+156.6
-160.2
+163.9
+153.9
+171.5
+177.8
+173.5
+176.5
-86.5
-73.5
-64.4
-62.5
-65.2
-60.5
2.3 ( 0.4
1100 ( 110
0.03 ( 0.01
5.3 ( 0.3
480 ( 50
4.2 ( 0.2
460 ( 30
0.54 ( 0.08
21 000 ( 3000f
1.6 ( 0.3
1,400 ( 200
.21 000f
1200 ( 100
.2300f
0.18 ( 0.03
750
D-Val-L-boroPro
2300 ( 200f
a We have avoided the (S) and (R) absolute configuration nomenclature here because of the confusing quirk that substitution of the -B(OH)2 group for
-CO2H in boronic acids results in a switch from the (S) to the (R) configuration and vice versa because of the Cahn-Ingold-Prelog priority rules.17 For
example, (S)-Ala-(S)-Pro is the stereochemical homolog of L-Ala-L-Pro, and (S)-Ala-(R)-boroPro is the stereochemical homolog of L-Ala-L-boroPro. b IC50
values obtained for the inhibitors equilibrated at pH 2 and pH 8 for times ranging from 18 to 24 h prior to addition to the enzyme. c Cyclization index, ratio
of IC50 values at pH 8 over those at pH 2. d Angles computed by HF/6-31G* method, with subsequent MP2/6-31G* refinement. e Angles computed by
HF/6-31G* method. f Inhibition most likely due to the presence of a minute amount of the active diastereomer.9
Scheme 1. Routes for the Preparation of 2 from N-Boc-L-Methionine (7)a and N-Boc-L-Homoserine (OBn; 9)b
a Reagents and conditions (route A): (i) L-boroAla-pn‚HCl salt, HATU, DIPEA, DMF; (ii) MeI; (iii) NaH, DMF, DCM, 0 °C; (iv) BCl3, DCM. bReagents
and conditions (route B): (i) L-boroAla-pn‚HCl salt, HATU, DIPEA, DMF; (ii) H2, Pd-C, EtOAc; (iii) MsCl, NEt3, DCM; (iv) LDA, THF, -78 °C; (v)
BCl3, DCM.
Scheme 2. Preparation of 1 from N-Boc-D-Methionine (11)a
Scheme 4. Possible Reaction Paths to Compounds 2 and 15a
a Reagents and conditions: (i) L-boroAla-pn‚HCl, HATU, DIPEA, DMF;
(ii) MeI; (iii) NaH, DMF, DCM, 0 °C; (iv) BCl3, DCM. Combined yield:
32% (1:2:12 ) 9:6:5).
Scheme 3. Synthesis of 1 and 2 Using Pf as N Protection
a Path A: amide N-alkylation. Path B: terminal N-alkylation. L ) leaving
group; P ) protective group.
Groupa
inhibition constant, is a measure of the enzyme-inhibitor
binding strength. It is much more difficult to measure than IC50,
owing to the slow, tight binding characteristics and the cycliza-
tion reaction of these inhibitors. We have previously pointed
out these difficulties and have developed new methods for
solving the problems.9 For these inhibitors, IC50 is valuable for
two reasons. First it provides a convenient way to measure their
cyclization tendency. Second, and perhaps more importantly,
the IC50 value is a better predictor of inhibitor effectiveness at
physiological pH.
a Reagents and conditions: (i) HCl, dioxane; (ii) PfBr, NEt3, Pb(NO3)2,
K3PO4, MeCN; (iii) MeI; (iv) NaH, DMF, DCM, 0 °C; (v) BCl3, DCM.
Inhibition of DPP IV by compounds 1-6 was evaluated and
the results summarized in Table 1. Results indicate that L-Ala-
L-boroPro (3) and L-Val-L-boroPro (5) are among the most
potent inhibitors of DPP IV, with subnanomolar Ki values. The
corresponding lactam, amino-L-lactam-L-boroAla (2), is less
potent by 3 to 4 orders of magnitude. Interestingly, however,
the opposite diastereomeric homolog, amino-D-lactam-L-boroAla
(1), inhibits some 500 times better, with a Ki value of 2.3 nM.
The CI values of compounds 3 and 5 are 2600 and 750,
respectively. In contrast, the IC50 values for the lactams 1 and
2 are pH-independent (CI ∼ 1), confirming that they do not
cyclize.
minutes was found optimal to give the slow-binding inhibitors
enough time to complex, yet not enough time to appreciably
shift the cis-trans equilibrium. The cyclization index (CI),
calculated by dividing the IC50 obtained at high pH by that at
low pH (IC50 (pH 8)/IC50 (pH 2)), was used as a means of
evaluating the cyclization tendency for each compound at
physiological pH. The greater the value of CI, the more the
equilibrium favors the cyclization.
In this paper, we report both IC50 and Ki values for each
inhibitor. The IC50 is the molar concentration of the inhibitor
that produces 50% inhibition of the enzyme activity. Ki, the

2394 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 10
Lai et al.
Table 2. Summary of Backbone Angles of Inhibitors Bound to DPP IV
by X-ray Crystallographic Determination of Structures of Complexes.
Multiple Entries Are Due to the Presence of Two or Four Liganded
Proteins Per Unit Cell.
ψ(P2)
(deg)
ω
(deg)
φ(P1)
(deg)
ID24
inhibitora
res.
ref.
1WCY L-Ile-L-Pro-L-Ile
133.9
129.4
179.6 -78.1 2.20 Å 25
179.7 -76.1
2AJB
2AJD
L-t-BuGly-L-Pro-L-Ile 134.7 -179.9 -80.0 2.75 Å 26
133.0 -179.9 -84.5
133.4 -179.9 -83.9
127.7
160.4 -179.9 -86.1 2.56 Å 26
180.0 -86.3
L-Pro-L-boroPro
169.9
162.2
154.4
179.8 -74.5
179.8 -78.6
180.0 -87.6
Figure 3. pH Dependence of lactam (1) vs L-Ala-L-boroPro (3):
unconstrained boronate L-Ala-L-boroPro (3, blue lines, squares) and
constrained lactam derivative amino-D-lactam-L-boroAla (1, red lines,
triangles). Solid lines and filled symbols are at pH 2.0; dashed lines
and open symbols are at pH 8.0.
a We have avoided the (S) and (R) absolute configuration nomenclature
here because of the confusing quirk that substitution of the -B(OH)2 group
for -CO2H in boronic acids results in a switch from the (S) to the (R)
configuration and vice versa because of the Cahn-Ingold-Prelog priority
rules.17 For example, (S)-Pro-(S)-Pro is the stereochemical homolog of L-Pro-
L-Pro, and (S)-Pro-(R)-boroPro is the stereochemical homolog of L-Pro-L-
boroPro.
Figure 3 compares enzyme inhibition curves of 1 and 3 as a
function of pH. It shows, as expected, that the inhibitory potency
of 3 is markedly pH dependent; the IC50 varies about 2,300-
fold, depending on whether 3 has been equilibrated at pH 2.0
or pH 8.0. The loss of inhibitory potency of 3 on equilibrating
at pH 8.0 is not irreversible, and can be fully recovered via 24
h in solution at pH 2.0. Compound 5 exhibits a similar, though
somewhat smaller, pH dependence with a CI of ∼750 (not
shown in Figure 3). In contrast, the inhibitory potency of lactams
1 (shown in Figure 3) and 2 (not shown) are entirely pH-
independent, thereby confirming our original expectation that
restrained structures would not cyclize. Note, while Figure 3
shows that 1 is clearly not as potent as 3 when 3 has been
equilibrated at pH 2.0, it also shows that 1 is actually more
potent than 3 at high pH values and, therefore, suggests that 1
could be more potent than 3 in vivo.
The calculated phi angles in Table 1 vary from -60.5° to
-86.5°. For comparison, the phi values of bound inhibitors in
Table 2 vary from -74.5° to -86.3°. Interestingly, of the
compounds listed in Table 1, phi of lactam 1 seems to be the
most optimally set for enzyme binding, which probably helps
to explain the potency of this constrained derivative. The other
inhibitors in Table 1 (3-6) have phi values of ∼-60°, which
is ∼15° to 20° away from optimal for enzyme binding. Note
that 3 and 5 are extremely potent enzyme inhibitors, while 4
and 6 are devoid of inhibitory potency, yet all have essentially
the same phi value. These differences are obviously determined
by the stereochemical configuration about P2. Thus, significance
of phi values for inhibitory potency cannot be assessed from
the current data.
The inhibitory activity of lactam 2 could have been caused
by a trace amount (ca. <1%) of 1 in the 2 sample; however,
What are the important contacts in the bound state of good
DPP IV inhibitors? As detailed in the X-ray crystallographic
studies of bound inhibitors,25,26 there are as many as three
H-bonds to the protonated N-terminal amino group of the P2
residue arising from Glu205, Glu206, and Tyr662. The carbonyl
oxygen of P2 is H-bonded to the carboxyamide group of Asn710
and guanidyl of Arg125. The P1 pyrollidine ring packs into the
S1 groove, attracted by hydrophobic forces. Only the L-L
diastereomer of the L-Pro-L,D-boroPro racemate binds to DPP
IV,26 showing the stereospecific contacts involving the N-
terminal nitrogen, the boron atom, and the pyrollidine ring. The
boron adduct to Ser630 is tetrahedral, with one of its hydroxyl
groups bound to the oxyanion hole (from Tyr547 and Ser631)
and the other bound to the catalytic His740.
Figure 4 shows our hypothetical scheme for rationalizing
X-ray data and inhibitor binding, showing the inhibitors from
their N-terminus along the CR--CdO bond of residue P2. The
dihedral angle between successive nitrogen atoms is the
backbone angle psi, whose approximate value is shown in each
subfigure. Figure 4a represents a composite of the X-ray
structures in Table 2, with five of the important H-bonds from
the enzyme to the ligand. Although there are no X-ray structures
of compounds 3 or 5, we expect that their backbones would
superimpose nicely on these structures, as shown in Figure 4a.
No X-ray structures exist yet for compounds 1, 2, 4, or 6 either,
but we can infer how H-bonding and steric repulsion might
influence the fit of these compounds and, thus, inhibitory
potency in the active site through the use of Figure 4b-d.
Figure 4a shows H-bonds donated from the enzyme to
inhibitors with P2 in the L configuration, which in solution are
1
none was detected by H NMR or LC-MS.
Computational Analysis. Molecular modeling was employed
to help ascertain from theory the conformational forms adopted
by compounds 1-6 (Table 1). Although psi and omega in
compounds 1 and 2 and phi in 3-6 are known approximately
from their ring constraints, for consistency all calculated values
of psi, omega, and phi are listed in Table 1. Gas-phase
calculations of the conformations of small molecule inhibitors
in solution have obvious limitations, but our sole aim here is to
seek any correlation of backbone conformation with IC50 values
for drug design purposes.
The six DPP IV inhibitors discussed here exhibit various
degrees of backbone conformational flexibility. For all com-
pounds, the calculated omega (ω) angles were not far from 180°,
producing a trans, planar amide bond relative to the opposing
R-carbons. In the lactams 1 and 2, omega is constrained by
inclusion in the closed rings to the trans conformer. Psi values
vary widely from about +150o for 3 and 5, to -160.5° for 2,
to about -160° in 4 and 6. From Table 1 it is clear that all
the highly effective inhibitors have psi angles in the +150°
neighborhood, regardless of the stereochemical configuration
about the P2 alpha carbon. This psi angle is similar to those
found for bound inhibitors in x-ray crystallographic studies,24-26
as shown in Table 2.
In the bound state, a trans (∼180o) torsion angle omega, or
ω (CR-C-N-CR), is required. Table 2 also shows the values
of psi, or ψ (N-CR-C-N), and phi, or φ (C-N-CR-C),
determined by X-ray crystallography of bound ligands.

Constrained Lactam Boronate DPP IV Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 10 2395
kon and koff rates should demonstrate that lactam 1 binds far
better than 4 and 6 because of its preformed psi as well as omega
angles as a result of the lactam ring constraints.
Conclusions
Conformationally restricted dipeptide boronates containing
a five-membered lactam ring (amino-lactam-L-boroAla) are
synthesized with L- and D-epimers. Among them, the D-epimer
(1) is found to be a very potent DPP IV inhibitor, exhibiting a
Ki value of 2.3 nM, whereas the L-epimer (2) showed only
moderate potency. The backbone constraint imposed by the
lactam peptidomimetic eliminates the N-B cyclization observed
in most dipeptide boronic acids. The biological activities of the
epimers are correlated with the greater similarity to the backbone
conformation of bound inhibitors, as revealed by X-ray crystal-
lographic studies of DPP IV complexes. The increased stability
should make these compounds effective at lower doses in animal
studies, with a longer duration of action. To our knowledge,
this is the first time that backbone-constrained peptidomimetic
inhibitors with low nanomolar potency for DPP IV have been
reported.
Figure 4. Schematic of proposed conformations of residue P2 for
unconstrained and lactam-constrained inhibitors bound to DPP IV active
site. The barrier or “wall” is comprised of the aromatic side chains of
Tyr666 and Phe357, as shown in X-ray structures.26
Experimental Section
Materials and Analysis. Reagents obtained from commercial
sources were used without further purification. NMR spectra were
recorded on a Bruker Avance 300 MHz spectrometer. Chemical
shifts were reported in parts per million (δ scale) relative to TMS
free to rotate about psi, namely, 3 and 5. A nominal psi value
of ∼140o is adopted, which is close to the average of the
inhibitor angles found in X-ray structures (Table 2). The Cd
OsNH group lies behind in the horizontal plane, roughly the
plane of the pyrollidine ring. The “wall” is formed by aromatic
side chains of Tyr666 and, perhaps to a lesser extent, Phe357.
There is plenty of room for the R groups in 3 and 5 (methyl
and i-propyl) in Figure 4a, accounting for their favorable binding
in Table 1. In addition, we have the fact that the t-Bu group in
X-ray structure 2AJB is easily accommodated, and in L-Pro-L-
boroPro (2AJD) the entire P2 proline ring lies in the region
shown by the R group, more or less perpendicular to the P1
proline ring. Figure 4b shows the steric repulsion between the
enzyme “wall” and the side chain R, which would exist in the
bound state for the D configuration of P2 when all other contacts
to the enzyme are formed. This argument can explain broadly
why 4 and 6 are such poor inhibitors.
Compound 2 is shown in Figure 4c as it might fit into the
active site with its boronate and carbonyl H-bonds to the enzyme
intact. Clearly, because of the lactam ring constraint, 2 cannot
form the N-terminal H bonds and, even with its otherwise
favorable L configuration, forms a poor inhibitor. Compound
1, depicted in Figure 4d can form all the correct H-bonds to
DPP IV, despite its unfavorable D configuration. The difference
in IC50 values between 1 and 3 could well be explained by some
steric repulsion to the lactam ring -CH2- group occurring in
1, as depicted in 4d. However, lactams 1 and 2, having no
pyrollidine ring, need not fit in the active site exactly the same
as compounds 3-6. This is suggested in Figure 4c,d by a slight
rotation of the lactams in such a way as to relieve steric
repulsion. Such a difference in binding could provide the needed
explanation for why lactam 1 is so much better an inhibitor
than 4 and 6, all possessing the D configuration.
1
(in CDCl3) or DSS (in D2O) for H and 13C NMR and boric acid
(in D2O) for 11B NMR. Mass spectra and HPLC retention times
were recorded on a Hewlett Packard HP LC/MSD system with UV
detector (monitoring at 215 nm), using a Discovery C18 569232-U
RP-HPLC column (12.5 cm × 4.6 mm, 5 µm) with solvent gradient
A ) water (0.1% TFA) and B ) acetonitrile (0.08% TFA) at 0.5
mL/min. Unless otherwise noted, all HPLC retention times are given
for an eluent gradient 2% B for the first 5 min, then from 2% to
98% B over 10 min, which was maintained for the next 10 min.
The crude targets were purified by RP-HPLC using a Varian
semipreparative system with a Discovery C18 569226-U RP-HPLC
column (25 cm × 21.2 mm, 5 µm) at 20 mL/min. HRMS were
performed by the Technical Services of the University of Michigan.
DPP IV Inhibition Assay. Dipeptidyl peptidase IV was purified
from human placenta.28 Assays were performed at ambient tem-
perature (23 °C) in 100 mM HEPES, pH 8.0, 0.14 M NaCl, with
the chromogenic substrate Ala-Pro-p-nitroanilide (Bachem). Ap-
proximately 1 nM of DPP IV enzyme was incubated with various
inhibitor concentrations, ranging between 10-4 and 10-11 M, for
10 min before starting the reaction by the addition of substrate to
a concentration of 30 µM. The absorbance at 410 nm was measured
30 min after the addition of substrate. The IC50 value is defined as
the concentration of inhibitor required to reduce the DPP IV activity
to 50% after a 10-minute incubation with the enzyme at 23 °C and
before the addition of the substrate. IC50 values were determined
by a nonlinear regression fit of the data to a sigmoidal dose-
response curve using the program Prism. DPP IV inhibition assays
were performed at 25 °C on a Molecular Devices SPECTRAmax
340PC384 microtiter plate reader, monitoring the absorbance at 410
nm and using H-AlaPro-pNA (Bachem) as a chromogenic substrate.
Inhibitor stock solutions (ca. 1 mg/mL) were prepared in either
0.01 N HCl solution, pH 2.0, or 0.1 M Hepes, 0.14 M NaCl buffer,
pH 8.0. All Xaa-boroPro inhibitors were allowed to stand for at
least 18 h to fully establish the cis-trans equilibrium before
assaying. Stock solutions were diluted with respective pH buffers
immediately prior to the commencement of the experiment.
Compounds 1 and 2 are reversible competitive inhibitors of DPP
IV. Ki values were determined by measuring the reaction rate versus
the substrate concentration at three different values of inhibitor
concentration. The data were fit by nonlinear regression to the
equation
The differences in IC50 values of 1 vs 4 and 6 must also arise
from the entropic advantage of the former. A flexible inhibitor
arriving at the active site preformed to the bound state
conformation will enjoy an entropic advantage due to higher
effective concentration and thus reflected in a higher “on rate”.27
As shown in Table 1, the theoretical values of psi have the
opposite sign (∼ -160o) for 4 and 6 in free solution instead of
the ∼ +140o presumeably required for binding. The relevant

2396 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 10
ν ) Vmax[S]/{[S] + Km(1 + [I]/Ki)}
Lai et al.
(1R)-1-[(3S)-3-Amino-2-oxopyrrolidin-1-yl]-ethylboronic Acid
Hydrochloride (2; Method A). Compound (8; 0.68 g, 1.5 mmol)
was dissolved in methyl iodide (10 mL) and stirred at room
temperature for 3 days. The reaction mixture was concentrated in
vacuum and evaporated three times from methylene chloride
solution. DMF-DCM (1:1; 16 mL) was added under Ar and cooled
to 0 °C. Sodium hydride (0.12 g of a 60% mineral oil suspension,
3 mmol) was added all at once, and the mixture was stirred at 0 °C
for an additional 6 h. NH4Cl (aq, 6.0 mL, half-saturated) was added
carefully to quench the reaction. The solution was concentrated in
vacuo to a small volume and partitioned between water (10 mL)
and EtOAc (20 mL). The aqueous phase was extracted with EtOAc
(2 × 20 mL). The combined organic phase was washed with aq
NaHCO3 and brine and then dried with MgSO4. After filtration
and concentration in vacuo, a pale yellow oil was yielded. Following
the general procedure for pinanediol and Boc removal, the target
compound (2) was obtained as a white powder (0.13 g, 41% yield).
using the program Prism (GraphPad Software, Inc.). Compounds
3 and 5 are slow, tight binding inhibitors. Measurement of Ki values
for these two compounds is further complicated by a slow
cyclization reaction that inactivates the inhibitors at the reaction
pH. This situation has been specifically addressed for DPP IV
inhibitors of this class by Gutheil and Bachovchin.9 Following the
same procedure, Ki values for compounds 3 and 5 were determined
by measuring the reaction rate versus the inhibitor concentration
at a fixed substrate concentration equal to 5 times the Km and the
lowest enzyme concentration that allows measurement of the rate
in 2 min. The resulting data were fit by nonlinear regression to the
simple equilibrium equation, as described by Gutheil and Bacho-
vchin.9 The substrate used for all Ki measurements was Ala-Pro-
para-nitroanilide, which has a Km value of 15 µM for DPP IV.
Initial rates were determined from a linear fit of the absorbance at
410 nm versus the time over 2 min on a Hewlet Packard 8453
UV/vis spectrophotometer.
Computations. Energy minimizations of these compounds were
conducted using BioMedCaChe (version 6.0, Fujitsu, Ltd.) and
Spartan (version 04, Wavefunction, Inc., Irvine, CA) programs.
First, using the molecular mechanics (MM3) or semiempirical
(AM1) method, a systematic search was performed by examining
all rotamers. The free-energy minimum structure for each compound
was extracted using HF/6-31G* calculations. For compounds 1 and
2, the structures were further optimized by ab initio calculations at
the MP2/6-31G* level. Backbone torsion angles ψ (P2), ω and φ
(P1) were obtained from the lowest-energy conformer (Table 1).
Circular Dichroism Spectroscopy. Measurements were carried
out on a JASCO J-810 CD spectropolarimeter at 25 °C, using a 1
mm quartz cell. Sample concentrations were about 0.25 mg/mL.
Samples at pH 8 were in 5 mM phosphate buffer, while pH 2
samples were in 0.01 M HCl.
1H NMR (D2O) δ 4.18 (t, J ) 9.4 Hz, H, H2NCH), 3.47-3.66
1
1
(m, 2H, CONCH2), 3.12 (q, J ) 7.5 Hz, H, BCHN), 2.57-2.67
(m, 1H, H2NCHCHACHB), 2.06-2.16 (m, 1H, H2NCHCHACHB),
1.20 (d, J ) 7.5 Hz, 2.85H, BCHCH3 of 2), 1.18 (d, J ) 7.5 Hz,
0.15H, BCHCH3 of 1). LC-MS (ESI+) for C6H13BN2O3 m/z (rel
intensity): 463.3 ((3 × (M - H2O) + H)+, 4), 309.2 ((2 × (M -
H2O) + H)+, 100), 155.3 ((M - H2O + H)+, 26); tr ) 6.0 min
(eluent gradient 0.5% B for the first 10 min and then from 0.5% to
20% B over 10 min); purity (2 and 5% 1) > 99%.
(1R)-1-[(3S)-3-Amino-2-oxopyrrolidin-1-yl]-ethylboronic Acid
Hydrochloride (2; Method B). Methanesulfonyl chloride (0.21 mL,
2.6 mmol) was added dropwise at 0 °C to a DCM solution (20
mL) containing 10 (0.55 g, 1.3 mmol) and NEt3 (0.72 mL, 5.2
mmol). The ice bath was removed after 30 min, and the mixture
was stirred for an additional 4 h at room temperature. A supple-
mentary 20 mL of DCM was added to the mixture, which was then
washed with water (10 mL), 0.1 N KHSO4 (10 mL), saturated aq
NaHCO3 (10 mL), and brine. The organic layer was then dried
with MgSO4, filtered, and concentrated in vacuo. The residue was
dissolved in anhydrous THF (10 mL), and LDA (0.6 mL, 2 M, 1.2
mmol) was added dropwise at -78 °C under Ar. The solution was
allowed to gradually warm to room temperature and continue to
react overnight. The reaction was quenched with saturated aq NH4-
Cl (5 mL) and then extracted with EtOAc (2 × 20 mL). The
combined organic phase was washed with aq NaHCO3 and brine
and dried with MgSO4. After filtration and concentration, a pale
yellow oil was yielded. Following the general procedure for
pinanediol and Boc removal, the target compound was obtained as
General Procedure for the Coupling Reaction. To a solution
of N-(tert-butoxycarbonyl)-L-methionine (7; 0.50 g, 2.0 mmol) in
DMF (8.0 mL) was sequentially added DIPEA (0.75 mL, 4.4
mmol), HATU (0.8 g, 2.2 mmol), and L-boroAla-pn‚HCl (550 mg,
2.1mmol) at 0 °C. The mixture was stirred at room temperature
overnight and concentrated in vacuum. The residue was redissolved
in EtOAc (50 mL) and washed with 0.1 N KHSO4 (3 × 15 mL),
saturated aq NaHCO3 (2 × 15 mL), and brine. The organic layer
was then dried with MgSO4, filtered, and condensed under vacuum.
General Procedure for Pinanediol and Boc Removal. The oily
protected crude (1.5 mmol) was dissolved in dry DCM (8.0 mL)
and cooled to -78 °C while BCl3 (1 M in DCM, 8.0 mL) was
added dropwise. The mixture was stirred at -78 °C for 1 h and
allowed to warm to room temperature for an additional 3-5 min.
The reaction mixture was then concentrated in vacuo and redis-
solved in ether (15 mL). Water (15 mL) was added and the aqueous
layer was washed twice with ether (2 × 15 mL). The aqueous layer
was concentrated in vacuo and further purified by semipreparative
RP-HPLC to afford the target compound as a white powder.
(1R)-1-[(3R)-3-Amino-2-oxopyrrolidin-1-yl]-ethylboronic Acid
Hydrochloride (1). The target compound was prepared from (14)
following the same procedure described in Method A. Compound
1 was obtained as a white powder after semipreparative RP-HPLC
1
a white powder (0.13 g, 46% yield). The H NMR spectrum was
the same as the one prepared by method A, except the content of
1 was varied from 5% to 20%.
(1R)-1-[(3S)-3-Amino-2-oxopyrrolidin-1-yl]-ethylboronic Acid
Hydrochloride (2; Method C). Following the same procedure in
the preparation of 1, the target compound was prepared from 8 as
a white powder after semipreparative RP-HPLC purification (16%
1
overall yield of the five steps). H NMR (D2O) δ 4.18 (t, J ) 9.4
1
Hz, H, H2NCH), 3.47-3.66 (m, 2H, CONCH2), 3.12 (q, J ) 7.5
Hz, 1H, BCHN), 2.57-2.68 (m, 1H, H2NCHCHACHB), 2.06-2.16
(m, 1H, H2NCHCHACHB), 1.20 (d, J ) 7.5 Hz, 2.85H, BCHCH3).
13C NMR (D2O) δ 172.29 (CO), 53.48 (H2NCH), 46.49 (CONCH2),
1
1
41.25 (BCHN), 26.32 (H2NCHCH2), 15.31 (BCHCH3). The H
purification (20% overall yield in three steps). H NMR (D2O) δ
NMR and 13C NMR assignments were consistent with the 1H-13C
HSQC. 11B NMR(D2O) δ 10.54 (br s). LC-MS (ESI+) for C6H13-
BN2O3 m/z (rel intensity): 463.3 ((3 × (M - H2O) + H)+, 6),
309.2 ((2 × (M - H2O) + H)+, 100), 155.3 ((M - H2O + H)+,
24); tr ) 6.0 min (eluent gradient 0.5% B for the first 10 min, then
from 0.5% to 20% B over 10 min); purity >99%. HRMS calcd for
C6H13BN2O3Na [M + Na]+, 195.0917; found, 195.0914.
1
4.17 (t, J ) 9.5 Hz, H, H2NCH), 3.47-3.68 (m, 2H, CONCH2),
3.13 (q, J ) 7.5 Hz, 1H, BCHN), 2.57-2.67 (m, 1H, H2-
1
NCHCHACHB), 2.04-2.17 (m, H, H2NCHCHACHB), 1.18 (d,
J ) 7.5 Hz, 3H, BCHCH3). 13C NMR (D2O) δ 172.30 (CO), 53.42
(H2NCH), 46.60 (CONCH2), 41.69 (BCHN), 26.26 (H2NCHCH2),
1
15.16 (BCHCH3). The H NMR and 13C NMR assignments were
1
consistent with the H-13C HSQC. 11B NMR (D2O) δ 10.38 (br
s). LC-MS (ESI+) for C6H13BN2O3 m/z (rel intensity): 463.3 ((3
× (M - H2O) + H)+, 3), 309.2 ((2 × (M - H2O) + H)+, 100),
155.3 ((M - H2O + H)+, 20); tr ) 6.0 min (eluent gradient 0.5%
B for the first 10 min, then from 0.5% to 20% B over 10 min);
purity >99%. HRMS calcd for C6H13BN2O3Na [M + Na]+,
195.0917; found, 195.0919.
(1R)-1-[N-(tert-Butoxycarbonyl)-L-methionyl]amino-ethylbo-
ronate (+)-Pinanediol Ester (8). N-(tert-butoxycarbonyl)-L-me-
thionine (7; 0.50 g, 2.0 mmol) was coupled to L-boroAla-pn‚HCl
(550 mg, 2.1 mmol) following the general procedure for the
coupling reaction. The target obtained was a white powder (0.80
g, 88% yield) after flash column chromatography (3/5 EtOAc/

Constrained Lactam Boronate DPP IV Inhibitors
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 10 2397
1
1
hexanes). H NMR(CDCl3) δ 6.46 (br s, H, BocNH), 5.24 (br d,
1H, CHCONH), 4.25-4.31 (m, 2H, BocNHCH, BOCH), 3.20-
3.24 (m, 1H, NHCHB), 2.11-2.56 (m, 4H, SCH2, pinane-H), 2.10
(s, 3H, SCH3), 1.85-2.09 (m, 5H, CH2CH2S, pinane-H), 1.44 (s,
9H, (CH3)3C), 1.39 (s, 3H, pinane-CH3), 1.29 (s, 3H, pinane-CH3),
1.28 (d, J ) 7.7 Hz, 1H, pinane-H), 1.22 (d, J ) 7.5 Hz, 3H, CH3-
CHNH), 0.84 (s, 3H, pinane-CH3). LC-MS (ESI+) for C22H39-
BN2O5S m/z (rel intensity): 477.3 ((M + Na)+, 20), 455.3 ((M +
H)+, 89), 399.2 ((M - CH2dC(CH3)2 + H)+, 67), 355.2 ((M -
CH2dC(CH3)2sCO2 + H)+, 20), 303.2 ((M - pinane - H2O +
H)+, 15), 247.1 ((M - CH2dC(CH3)2spinane - H2O + H)+, 100),
203.1 ((M - CH2dC(CH3)2sCO2spinane - H2O + H)+, 15); tr
) 18.3 min; purity >99%. HRMS calcd for C22H40BN2O5S [M +
H]+, 455.2751; found, 455.2756.
was dissolved in a 1,4-dioxane solution of HCl (5.0 mL, 4 M) in
an ice-cooled bath and stirred for 3 h at room temperature. After
being concentrated in vacuo and evaporated three times from
chloroform solution, the residue was suspended in anhydrous
acetonitrile (10 mL) and sequentially added Pb(NO3)2 (0.66 g, 2.0
mmol), K3PO4 (0.85 g, 4.0 mmol), and a solution of PfBr (0.80 g,
2.5 mmol) in CH3CN (3.0 mL) at room temperature. After being
stirred for 48 h, the reaction mixture was filtered through a pad of
Celite, and the residue was thoroughly washed with chloroform.
The combined filtrates and washings were evaporated and then
partitioned between 2.5% aq citric acid (10 mL) and ether (40 mL).
The dried organic phase was evaporated to yield the crude product,
which was purified by flash column chromatography using EtOAc/
hexanes (1/4) to afford 14 (0.94 g, 79% yield) as a white powder.
1H NMR (CDCl3) δ 7.62-7.67 (m, 2H, Pf-H), 7.19-7.40 (m, 11H,
(1R)-1-[N-(tert-Butoxycarbonyl)-L-homoseryl]amino-ethylbo-
ronate (+)-Pinanediol Ester (10). N-(tert-Butoxycarbonyl)-O-
benzyl-L-homoserine (9; 0.62 g, 2 mmol) was coupled to L-boroAla-
pn‚HCl (550 mg, 2.1 mmol) following the general procedure for
coupling reaction except using DCM (15 mL) as solvent. After
concentration, the residue was redissolved in EtOAc (15 mL), and
10% Pd/C (100 mg) was added. The mixture was stirred under a
hydrogen atmosphere (50 psi) for 5 h. The resulting suspension
was filtered through Celite, and the solvent was evaporated. After
purification by flash column chromatography using EtOAc/hexanes
(4/5), the target was obtained as a white powder (0.45 g, 53% yield).
1H NMR (CDCl3) δ 6.57 (br d, 1H, exchangeable with D2O,
BocNH), 5.57 (br d, 1H, exchangeable with D2O, CHCONH), 4.12-
4.33 (m, 2H, BocNHCH, BOCH), 3.69-3.73 (m, 2H, CH2OH),
3.24-3.30 (m, H, NHCHB), 2.85 (br s, H, exchangeable with
D2O, CH2OH), 1.80-2.35 (m, 7H, CH2CH2OH, pinane-H), 1.45
(s, 9H, (CH3)3C), 1.41 (s, 3H, pinane-CH3), 1.29 (s, 3H, pinane-
CH3), 1.23(d, J ) 7.4 Hz, 3H, CH3CHNH), 0.85 (s, 3H, pinane-
CH3). LC-MS (ESI+) for C21H37BN2O6 m/z (rel intensity): 447.2
((M + Na)+, 11), 425.2 ((M + H)+, 100), 369.2 ((M - CH2d
C(CH3)2 + H)+, 69), 325.3 ((M - CH2dC(CH3)2 - CO2 + H)+,
21), 273.2 ((Mspinane - H2O + H)+, 17), 217.2 ((M - CH2d
C(CH3)2spinane - H2O + H)+, 69), 173.3 ((M - CH2dC(CH3)2
- CO2 - pinane - H2O + H)+, 10); tr ) 15.8 min; purity >97%.
(1R)-1-(1-Aminocyclopropane-carbonyl)-ethylboronic Acid
Hydrochloride (12). The target was prepared from N-(tert-
butoxycarbonyl)-D-methionine (11) following the same procedure
to prepare (2; method A). After semipreparative RP-HPLC purifica-
tion, the target was afforded as a white powder from (8% overall
yield of the four steps). 1H NMR (D2O) δ 2.78 (q, J ) 7.4 Hz, 1H,
BCHN), 1.57-1.59 (m, 4H, cyclopropane-H), 1.09 (d, J ) 7.4 Hz,
3H, BCHCH3). LC-MS (ESI+) for C6H13BN2O3 m/z (rel inten-
sity): 485.3 ((3 × (M - H2O) + Na)+, 3), 463.3 ((3 × (M -
H2O) + H)+, 31), 309.2 ((2 × (M - H2O) + H)+, 100), 155.3 ((M
- H2O + H)+, 31); tr ) 4.3 min (eluent gradient 0.5% B for the
first 10 min, then from 0.5% to 20% B over 10 min); purity >95%.
HRMS calcd for C6H13BN2O3Na [M + Na]+, 195.0917; found,
195.0925.
1
1
Pf-H), 6.69 (br d, H, CONH), 4.31 (dd, J ) 8.8, 2.0 Hz, H,
1
1
BOCH), 3.97-4.00 (m, H, PfNHCH), 2.90-2.92 (m, H, NH-
1
CHB), 2.85 (br d, H, PfNH), 1.95-2.65 (m, 6H, SCH2, pinane-
H), 1.93 (s, 3H, SCH3), 1.65-1.91 (m, 3H, CH2CH2S, pinane-H),
1.38 (s, 3H, pinane-CH3), 0.93-1.30 (m, 7H, CH3CHNH, pinane-
CH3, pinane-H), 0.85 (s, 3H, pinane-CH3). LC-MS (ESI+) for
C36H43BN2O3S m/z (rel intensity): 595.5 ((M + H)+, 42), 461.2
((M-pinane + H)+, 3), 241.2 (C19H13+, 100); tr ) 19.4 min; purity
>95%. HRMS calcd for C36H44BN2O3S [M + H]+, 595.3166;
found, 595.3179.
(1R)-1-[(S)-Azetidine-2-carbonyl]amino-ethylboronic Acid (15).
L-Azetidine-2-carboxylic acid (0.1 g, 1.0 mmol) was Boc-protected
at the N-terminus following the standard procedure.29 The protected
product was then coupled to L-boroAla-pn‚HCl (270 mg, 1.1 mmol)
following the general procedure for the coupling reaction. The
protection groups were subsequently removed following the general
procedure for pinanediol and Boc removal to give the target as a
1
1
1
white powder (0.12 g, 57% yield). H NMR (D2O) δ 5.10 (t, J )
8.4 Hz, 1H, HNCHCO), 3.98-4.22 (m, 2H, NHCH2), 3.01 (q, J )
7.6 Hz, 1H, BCHN), 2.76-2.85 (m, 1H, NHCH2CHACHB), 2.59-
2.67 (m, 1H, NHCH2CHACHB), 1.18 (d, J ) 7.6 Hz, 3H,
BCHCH3). LC-MS (ESI+) for C6H13BN2O3 m/z (rel intensity):
463.3 ((3 × (M - H2O) + H)+, 10), 309.2 ((2 × (M - H2O) +
H)+, 100), 155.3 ((M - H2O + H)+, 38); tr ) 4.9 min; purity
>95%. HRMS calcd for C6H13BN2O3Na [M + Na]+, 195.0917;
found, 195.0912.
Acknowledgment. Financial support from Arisaph Pharm-
ceuticals, Inc. is gratefully acknowledged. One of the authors,
G.D.H., was supported by NIH Grant GM067985. We also thank
Dr. James Robertson for proofreading the manuscript.
Supporting Information Available: 1H NMR spectra for the
1
target compounds 1, 2, 12, and 15; 13C NMR, 11B NMR, and H/
13C HSQC spectra for 1 and 2; NOESY spectrum for 2; and CD
spectra of compounds 1 and 2 at low and high pH. This material
is available free of charge via the Internet at http://pubs.acs.org.
(1R)-1-[N-(tert-Butoxycarbonyl)-D-methionyl]amino-ethylbo-
ronate (+)-Pinanediol Ester (13). The target compound was
prepared from N-(tert-butoxycarbonyl)-D-methionine (11) following
the same procedure to prepare (8). 1H NMR (CDCl3) δ 6.63 (br s,
1H, BocNH), 5.30 (br d, 1H, CHCONH), 4.24-4.29 (m, 2H,
BocNHCH, BOCH), 3.08-3.12 (m, 1H, NHCHB), 2.10-2.53 (m,
4H, SCH2, pinane-H), 2.07 (s, 3H, SCH3), 1.83-2.06 (m, 5H, CH2-
CH2S, pinane-H), 1.42 (s, 9H, (CH3)3C), 1.37 (s, 3H, pinane-CH3),
1.27 (s, 3H, pinane-CH3), 1.26 (d, J ) 7.7 Hz, 1H, pinane-H), 1.18
(d, J ) 7.5 Hz, 3H, CH3CHNH), 0.82 (s, 3H, pinane-CH3). LC-
MS (ESI+) for C22H39BN2O5S m/z (rel intensity): 477.3 ((M +
Na)+, 23), 455.3 ((M + H)+, 100), 399.3 ((M - CH2dC(CH3)2 +
H)+, 23), 355.4 ((M - CH2dC(CH3)2 - CO2 + H)+, 5), 303.4
((Mspinane - H2O + H)+, 5), 247.3 ((M - CH2dC(CH3)2s
pinane - H2O + H)+, 35), 203.4 ((M - CH2dC(CH3)2 - CO2 -
pinane - H2O + H)+, 3); tr ) 18.3 min; purity >98%. HRMS
calcd for C22H40BN2O5S [M + H]+, 455.2751; found, 455.2741.
(1R)-1-[N-(9-Phenylfluoren-9-yl)-D-methionyl]amino-ethylbo-
ronate (+)-Pinanediol Ester (14). Compound (13; 0.91 g, 2 mmol)
References
(1) De Meester, I.; Korom, S.; Van Damme, J.; Scharpe, S. CD26, let it
cut or cut it down. Immunol. Today 1999, 20, 367-375.
(2) von Bonin, A.; Huhn, J.; Fleischer, B. Dipeptidyl-peptidase IV/CD26
on T cells: Analysis of an alternative T-cell activation pathway.
Immunol. ReV. 1998, 161, 43-53.
(3) Hong, W. J.; Doyle, D. Molecular dissection of the NH2-terminal
signal/anchor sequence of rat dipeptidyl peptidase IV. J. Cell Biol.
1990, 111, 323-328. Iwaki-Egawa, S.; Watanabe, Y.; Kikuya, Y.;
Fujimoto, Y. Dipeptidyl peptidase IV from human serum: purifica-
tion, characterization, and N-terminal amino acid sequence. J.
Biochem. 1998, 124, 428-433. Durinx, C.; Lambeir, A. M.; Bosmans,
E.; Falmagne, J. B.; Berghmans, R.; Haemers, A.; Scharpe, S.; De
Meester, I. Molecular characterization of dipeptidyl peptidase activity
in serum: soluble CD26/dipeptidyl peptidase IV is responsible for
the release of X-Pro dipeptides. Eur. J. Biochem. 2000, 267, 5608-
5613.
(4) Leitung, B.; Pryor, K. D.; Wu, J. K.; Marsilio, F.; Patel, R. A.; Craik,
C. S.; Ellman, J. A.; Cummings, R. T.; Thornberry, N. A. Catalytic
properties and inhibition of proline-specific dipeptidyl peptidases II,
IV and VII. Biochem. J. 2003, 371, 525-532.

2398 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 10
Lai et al.
(5) Meintlein, R. Dipeptidyl-peptidase IV (CD 26): Role in the inactiva-
tion of regulatory peptides. Regul. Pept. 1999, 85, 9-24.
(6) Deacon, C. F. Therapeutic strategies based on glucagon-like peptide
1. Diabetes 2004, 53, 2181-2189.
(7) Drucker, D. J. Therapeutic potential of dipeptidyl peptidase IV
inhibitors for the treatment of type 2 diabetes. Expert Opin. InVest.
Drugs 2003, 12, 87-100. Augustyns, K.; Van der Veken, P.; Senten,
K.; Haemers, A. Dipeptidyl peptidase IV inhibitors as new therapeutic
agents for the treatment of type 2 diabetes. Expert Opin. Ther. Pat.
2003, 13, 499-510.
(8) Flentke, G. R.; Munoz, E.; Huber, B. T.; Plaut, A. G.; Kettner, C.
A.; Bachovchin, W. W. Inhibition of dipeptidyl aminopeptidase IV
(DP-IV) by Xaa-boroPro dipeptides and use of these inhibitors to
examine the role of DP-IV in T-cell function. Proc. Natl. Acad. Sci.
U.S.A. 1991, 88, 1556-1559.
(9) Gutheil, W. G.; Bachovchin, W. W. Separation of L-Pro-DL-boroPro
and its component diastereomers and kinetic analysis of their
inhibition of dipeptidyl peptidase IV. A new method for the analysis
of slow, tight-binding inhibition. Biochemistry 1993, 32, 8723-8931.
(10) Snow, R. J.; Bachovchin, W. W. Boronic acid inhibitors of dipeptidyl
peptidase IV: A new class of immunosuppressive agents. AdV. Med.
Chem. 1995, 3, 149-177.
(11) Sudmeier, J. L.; Gunther, U. L.; Gutheil, W. G.; Coutts, S. J.; Snow,
R. J.; Barton, R. W.; Bachovchin, W. W. Solution structures of active
and inactive forms of the DPP IV (CD26) inhibitor Pro-boroPro
determined by NMR spectroscopy. Biochemistry 1994, 33, 12427-
12438.
(12) Snow, R. J.; Bachovchin, W. W.; Barton, R. W.; Cambell, S. J.;
Coutts, S. J.; Freeman, D. M.; Gutheil, W. G.; Kelly, T. A.; Kennedy,
C. A.; Krolikowski, D. A.; Leonard, S. F.; Pargellis, C. A.; Tong,
L.; Adams, J. Studies on proline boronic acid dipeptide inhibitors of
dipeptidyl peptidase IV: Identification of a cyclic species containing
a B-N bond. J. Am. Chem. Soc. 1994, 116, 10860-10869.
(13) Coutts, S. J.; Kelly, T. A.; Snow, R. J.; Kennedy, C. A.; Barton, R.
W.; Adams, J.; Krolikowski, D. A.; Freeman, D. M.; Campbell, S.
J.; Ksiazek, J. F.; Bachovchin, W. W. Structure-activity relationships
of boronic acid inhibitors of dipeptidyl peptidase IV. 1. Variation of
the P2 position of Xaa-boroPro dipeptides. J. Med. Chem. 1996, 39,
2087-2094.
(14) Heins, J.; Welker, P.; Schoenlein, C.; Born, I.; Hartrodt, B.; Neubert,
K.; Tsuru, D.; Barth, A. Mechanism of proline-specific proteinases:
(I) Substrate specificity of dipeptidyl peptidase IV from pig kidney
and proline-specific endopeptidase from Flavobacterium menin-
gosepticum. Biochim. Biophys. Acta. 1988, 954, 161-169.
(15) Hu, Y.; Ma, L.; Wu, M.; Wong, M. S.; Li, B.; Corral, S.; Yu, Z.;
Nomanbhoy, T.; Alemayehu, S.; Fuller, S. R.; Rosenblum, J. S.;
Rozenkrants, N.; Minimo, L. C.; Ripka, W. C.; Szardenings, A. K.;
Kozarich, J. W.; Shreder, K. R. Synthesis and structure-activity
relationship of N-alkyl Gly-boro-Pro inhibitors of DPP4, FAP, and
DPP7. Bioorg. Med. Chem. Lett. 2005, 15, 4239-4242.
(16) Freidinger, R. M.; Perlow, D. S.; Veber, D. F. Protected lactam-
bridged dipeptides for use as conformational constraints in peptides.
J. Org. Chem. 1982, 47, 104-109. Freidinger, R. M. Design and
synthesis of novel bioactive peptides and peptidomimetics. J. Med.
Chem. 2003, 46, 5553-5566.
(17) Cahn, R. S.; Ingold, C.; Prelog, V. Specification of molecular chirality.
Angew. Chem., Int. Ed. Engl. 1966, 5, 385-415.
(18) Pechenov, A.; Stefanova, M. E.; Nicholas, R. A.; Peddi, S.; Gutheil,
W. G. Potential transition state analogue inhibitors for the penicillin-
binding proteins. Biochemistry 2003, 42, 579-588.
(19) Kienhofer, A. 1-Hydroxy-7-azabenzotriazole (HOAt) and N-[(dim-
ethylamino)-1H-1,2,3-triazolo-[4,5-b]pyridin-1-yl-methylene]-N-me-
thylmethanaminium hexafluorophosphate N-oxide (HATU). Synlett
2001, 11, 1811-1812.
(20) Thaisrivongs, S.; Pals, D. T.; Turner, S. R.; Kroll, L. T. Conforma-
tionally constrained renin inhibitory peptides: γ-Lactam-bridged
dipeptide isostere as conformational restriction. J. Med. Chem. 1988,
31, 1369-1376.
(21) Matteson, D. S.; Lu, J. Asymmetric synthesis of 1-acyl-3,4-
disubstituted pyrrolidine-2-boronic acid derivatives. Tetrahedron:
Asymmetry 1998, 9, 2423-2436.
(22) Lubell, W. D.; Rapoport, H. Configurational stability of N-protected
R-amino aldehydes. J. Am. Chem. Soc. 1987, 109, 236-239.
(23) Feldman, P. L.; Rapoport, H. Synthesis of optically pure ∆4-
tetrahydroquinolinic acids and hexahydroindolo[2,3-a]quinolizines
from L-aspartic acid. Racemization on the route to vindoline. J. Org.
Chem. 1986, 51, 3882-3890.
(24) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J. B.; Meyer, E. F.;
Brice, M. D.; Kennard, O.; Shimanouchi, T.; Tasumi, M. The Protein
Data Bank: A computer-based archival file for macromolecular
structures. J. Mol. Biol. 1977, 112, 535.
(25) Hiramatsu, H.; Yamamoto, A.; Kyono, K.; Higashiyama, Y.; Fuku-
shima, C.; Shima, H.; Sugiyama, S.; Inaka, K.; Shimizu, R. The
crystal structure of human dipeptidyl peptidase IV (DPPIV) complex
with diprotin A. Biol. Chem. 2004, 385, 561-564.
(26) Engel, M.; Hoffmann, T.; Manhart, S.; Heiser, U.; Chambre, S.;
Huber, R.; Demuth, H. U.; Bode, W. Rigidity and flexibility of
dipeptidyl peptidase IV: Crystal structures of and docking experi-
ments with DPIV. J. Mol. Biol. 2006, 355, 768-783.
(27) Khan, A. R.; Parrish, J. C.; Fraser, M. E.; Smith, W. W.; Bartlett, P.
A.; James, M. N. G. Lowering the entropic barrier for binding
conformationally flexible inhibitors to enzymes. Biochemistry 1998,
37, 16839-16845.
(28) (a) Pu¨schel, G.; Mentlein, R.; Heymann, E. Isolation and character-
ization of dipeptidyl peptidase IV from human placenta. Eur. J.
Biochem. 1982, 126, 359-365. (b) Heins, J.; Welker, P.; Scho¨nlein,
C.; Born, I.; Hartrodt, B.; Neubert, K.; Tsuru, D.; Barth, A.
Mechanism of proline-specific proteinases: (I) Substrate specificity
of dipeptidyl peptidase IV from pig kidney and proline-specific
endopeptidase from FlaVobacterium meningosepticum. Biochim.
Biophys. Acta 1988, 954, 161-169. (c) Rahfeld, J.; Schutkowski,
M.; Faust, J.; Neubert, K.; Barth, A.; Heins, J. Extended investigation
of the substrate specificity of dipeptidyl peptidase IV from pig kidney.
Biol. Chem. Hoppe-Seyler 1991, 372, 313-318.
(29) Guanti, G.; Riva, R. Synthesis of chiral non-racemic azetidines by
lipase-catalyzed acetylations and their transformation into amino
alcohols: precursors of chiral catalysts. Tetrahedron: Asymmetry
2001, 12, 605-618.
JM061321+